britt session vii tuesday (variability) small file
TRANSCRIPT
The Snap Sampler tm
Sampling VariabilitySampling Variabilityand Trend Monitoringand Trend Monitoring
North American Environmental Field
ConferenceTampa, Florida
January 15, 2007
Mr. Sandy Britt, PG, CHGProHydro, Inc.
Fairport, New [email protected]
Fewer Variables with SNAP SAMPLER
30
35
40
45
50
55
60
65
70
0 4 8 12 16 20 24
Sampling events
Con
cent
ratio
n
Many Variables with Purge and Pour
30
35
40
45
50
55
60
65
70
0 4 8 12 16 20 24
Sampling events
Con
cent
ratio
n
Low noise data
High noise data
What are sources of data error and variability?
Laboratory?Hold timeCross-contaminationManual vs. autosampler
Collection Method?Bail purge and samplePump purge and sampleLow flow purge and samplePassive (noPassive (no--purge) samplingpurge) sampling
Sample Handling?Bottle fillingTransport
Dye source with gravity feed to injection port
Simulated 4-inch well
Upgradient/ influent reservoir
Piezometers
Dye port
Constant head reservoir
Homer’s water supply bucket
Influent supply from constant head reservoir Effluent drain
Effluent reservoir
DTSC Sand Tank Well Model“homogeneous” model well
Britt, SL, 2005, Testing the In-Well Horizontal Laminar Flow Assumption with a Sand Tank Well Model. Ground Water Monitoring and Remediation 25, no. 3 p.73-81
percent of initial dye concentration
33%
40%
35%
Flow weighted averaging
Flow-weighted averaging effect
∇ Water chemistry changes as a well is pumped
∇ Why does chemistry change?
∇ “Stagnant” water?
Or…
∇ A varying mix of water entering the pump?
Active (Purge) Sampling Methods
Varljen, et al., 2006, Numerical Simulations to Assess the Monitoring Zone Achieved during Low-Flow Purging and Sampling, GWMR, 26: p. 44-52
Analysis of
Steady-State
purging
Hydraulics controls flow, including water coming from beyond the screen zone
But purge time controls what water discharges from the pump
Traditional and Low Flow Purging Static vs. Dynamic
Stability parameters monitored…
Achieving “stability” depends on:
∇ Pre-purge well stratification∇ Stratification in the aquifer∇ Inflow characteristics∇ Shadow effects in the aquifer∇ Density effects
∇ Stability criteria∇ Barn door (e.g. 10%)∇ Reasonable measures (EC or
ORP +/- 10 units, without directional drift)
∇ How often you measure
Purge Equilibration
This is what “Stability”looks like
How long does purge equilibration take?
Achieving true “stability” (i.e. flow-weighted-average)depends on:
∇ Well diameter∇ Well length∇ Pump position∇ Contaminant position∇ Other heterogeneity
Quick Calculation:4” well, 10 foot screen, pump in the middle,
250 ml/min purge rate
(2.5L/ft)*5 ft/(0.25L/min) = 50 min (plug flow)
>50 minutes
Not too long????
What about Sampling Passively?
What is Passive Sampling?
Sampling without purgingSampling without purging>Deploy in advance>Deploy in advance>Relies on >Relies on ““passivepassive”” flushingflushing
Examples:Polyethylene Diffusion Bag sampler (USGS)Rigid Porous Pipe sampler (USGS)Dialysis Membrane Sampler (USGS)Gore Sorber (WL Gore)
Snap Sampler (ProHydro)
• Natural flow delivered to well
• Ambient (passive) mixing according to native flow dynamics
• Stratification testing?
• Sample from same position(the key to consistency)
Passive EquilibrationCan Limit Variables
The Snap Sampler is a dedicatedpassive sampling system
• Deploy double-ended bottles in an open position.
• Sample after short or long residence time in the well—1-2 weeks or 3 or 6 months
• Mechanical or electric trigger closes bottles in situ.
• Sample transfer is not requiredSample transfer is not requiredat the well head for VOCs-No exposure to air
How the Snap Sampler works….
Insert
40ml 125ml
Set Snap Cap
• Load & Set Snap Caps
How the Snap Sampler works…continued
Lower downhole
• Modular samplers allow up to 4 bottles per trigger
• Multiple triggers can be used for multiple sampling depths
Hang on Dock Ring Secure
• Mechanical or electric trigger
Attach Trigger
Seal in situ,
reduce surface handling
In Situ:
• Sample at the same position each sampling event
• Sample collection takes place submerged in the well
No well-head sample transfer required
• Very good correlations
• VOCS, 1,4-dioxane, anionsParsons, 2005, McClellan Air Force Base, Sacramento, CA
Example data
Spotting data trends may depend on consistency of your collection method…
Many variables for purge:Many variables for purge:actual concentration, pump position, pumping duration, stability criteria requirements, wind, temperature, site surface or ambient air contamination, bottle fill rate, pour technique, speed of bottle closure, filtration, transport, analytical variability.
Fewer variables Fewer variables with with in situ sealedin situ sealedsamples:samples: actual concentration, transport, analytical variability.
Interpretation of simple data trendssimple data trends is easier with less random error
Illustration only, not site data
Simple Declining Trend
20
25
30
35
40
45
50
55
60
65
70
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sampling event
Con
cent
ratio
n
"Actual"
30% Random
2025303540455055606570
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sampling event
Con
cent
ratio
n
10% Random
2025303540455055606570
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sampling event
Con
cent
ratio
n
Interpretation of more complicatedmore complicated data trends is easier with less random error
Illustration only, not site data
Declining then Increasing Trend
2025303540455055606570
1 3 5 7 9 11 13 15 17 19 21 23
Sampling event
Con
cent
ratio
n"Actual"
10% Random
2025303540455055606570
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sampling event
Con
cent
ratio
n
30% Random
2025303540455055606570
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sampling event
Con
cent
ratio
n
Interpretation of very complicatedvery complicated data trends is easier with less random error
Illustration only, not site data
Complex Trend
2025303540455055606570
1 3 5 7 9 11 13 15 17 19 21 23
Sampling event
Con
cent
ratio
nActual
10% Random
2025303540455055606570
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sampling event
Con
cent
ratio
n
30% Random
2025303540455055606570
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sampling event
Con
cent
ratio
n
Example Data Variation
0
20
40
60
80
100
120
140
160
Sampling Quarter
TCE
mic
rogr
ams
per l
iter
purge 24.3 134 54.9 120 28.9 92.7 71 5.6 12.6
1 2 3 4 5 6 7 8 9
Example Data Variation
0
20
40
60
80
100
120
140
160
Sampling Quarter
TCE
mic
rogr
ams
per l
iter
Snap 56.9 72.6 64.1 103.0 39.7 65.6 40.9 50.8 50.4
1 2 3 4 5 6 7 8 9
TCE Example, site in Southern California
Range: 5.6 to 134Avg. RPD Q to Q: 100%Median RPD Q to Q: 94%
Private site, unpublished data
Range: 39.7 to 103Avg. RPD Q to Q: 36%Median RPD Q to Q: 35%
Purge Sealed in situ
Pump purge and bail sample Passive Snap Sample
TCE Example, Quarter to Quarter change
Example Data Variation
0
20
40
60
80
100
120
140
160
Sampling Quarter
TCE
mic
rogr
ams
per l
iter
Snap 56.9 72.6 64.1 103.0 39.7 65.6 40.9 50.8 50.4
purge 24.3 134 54.9 120 28.9 92.7 71 5.6 12.6
1 2 3 4 5 6 7 8 9
Relative Percent DifferenceQuarter to following quarter
0%
20%
40%
60%
80%
100%
120%
140%
160%
180%
1 2 3 4 5 6 7 8
Event
RPD
Snap Sample Pump Purge
TCE data superimposed Quarter to Quarter Relative Change
• Directional dynamic unchanged
• Quarterly concentration change less exaggerated
Overall statistics indicate differences in methods:
Purgen = 100 comparison pairs
Mean RPD Q to Q: 66%
Median RPD Q to Q: 51%
Mean % change Q-Q: 298%Median % change: Q-Q: 71%
Sealed in situn = 81 comparison pairs
Mean RPD Q to Q: 48%
Median RPD Q to Q: 37%
Mean % change Q-Q: 138%Median % change: Q-Q: 55%
Note: differences include the actual changes in
concentration…
• Reduced variation possible through consistent downhole passive sampling method
• In Situ sealed samples avoid error from surface handling
www.SnapSampler.com
Harbec Plastics Ontario, New York
“Technical Innovation with Environmental Responsibility”
• Passive method adds consistency by avoiding variables introduced during purge step
• Data trend more closely reflects downhole condition